Low-resolution Face Recognition Research Articles

Degradation model and attention guided distillation approach for low resolution face recognition

Deep convolution neural networks (CNN) have shown their efficacy in face recognition tasks due to their ability to extract highly discriminant face representations from face images. On high-resolution benchmark datasets, outstanding identification and verification results have been achieved. However, the performance of these algorithms is significantly degraded when tested on low-resolution (LR) images such as those captured from surveillance cameras. A straightforward solution to this problem is to use both high-resolution (HR) images and corresponding down-sampled LR images during training. Although this strategy improves the performance for LR images, it has some limitations. First, there is a significant difference between down-sampled LR images and LR images from surveillance cameras, leading to performance saturation at an earlier stage. Another limitation is the deterioration in the performance of HR images. In this work, solutions to both these limitations are proposed. A degradation model is proposed that synthesizes LR images from corresponding HR, emulating the real-world degradation effects in synthetic data, thus enabling the face recognition system to tolerate various blurry and noisy effects. To address the deterioration in the performance of HR images, an attention-guided distillation is proposed, which utilizes attention maps from convolutional layers in combination with deep features to transfer informative HR features from teacher to student network. The attention maps from the teacher network guide the student network to a better optimum and produce resolution robust face representations. The results of the proposed approach on the popular LR datasets like SCface, Coxface, and PaSC show that it outperforms the recent state-of-the-art (SOTA) techniques by a significant margin demonstrating its effectiveness for different cross-resolution scenarios.

Seg-DGDNet: Segmentation Based Disguise Guided Dropout Network for Low Resolution Face Recognition

Seg-DGDNet: Segmentation Based Disguise Guided Dropout Network for Low Resolution Face Recognition

Coupled discriminative manifold alignment for low-resolution face recognition

In practical applications, due to a long distance between the monitored population and monitoring equipment, the face images or human pose captured by the cameras often incur low-resolution (LR), small size, and poor quality, which leads to extreme difficulty in directly matching an LR face with the high-resolution (HR) ones in the gallery. In this paper, we propose a novel coupled discriminative manifold alignment (CDMA) method for LR face recognition. Specifically, in the training stage the principal component analysis (PCA) is first used to reduce the dimensional gap between LR and HR facial features. Next the LR face images and the corresponding HR face images are converted into a common shared feature subspace by learning two linear mappings in a supervised manner, where the neighborhood samples within the same class and from different classes are jointly exploited to align the manifold structures of LR and HR faces. In the test stage, for a given LR face in the probe set, two learned coupled mappings (CMs) are applied to match the HR images in the gallery set through the correlative metric. Thorough experimental results on three representative face databases verify the effectiveness of the proposed method in comparing with other state-of-the-art competitors. In particular, the proposed method is capable of yielding more competitive recognition performance than other predecessors when lower dimensional feature subspaces are applied to match the expected HR faces.

Low-resolution Face Recognition and Sports Training Action Analysis based on Wireless Sensors

Low-resolution Face Recognition and Sports Training Action Analysis based on Wireless Sensors

Demographic Bias in Low-Resolution Deep Face Recognition in the Wild

Face biometrics play a primary role in smart cities, from consumer- to organizational-level applications. This class of technologies has been recently shown to emphasize performance disparities across gender and ethnic groups. Prior work on demographic bias in deep face recognition has tended to focus exclusively on high-resolution images, leaving out low-resolution images captured, for example, by a certain surveillance camera at a distance. Notable reasons for this focus include the lack of low-resolution face image data sets that report user's demographic attributes. Therefore, demographic bias in low-resolution deep face recognition in the wild still remains under-explored. In this paper, we propose a framework for exploring demographic disparities in low-resolution face recognition at a distance. To this end, we devised a deep generative approach to turn high-resolution face images to their low-resolution counterpart. We then trained state-of-the-art face recognition models on different combinations of high- and low-resolution images. Finally, we assessed the model demographic disparities on artificially degraded images from five data sets. Our results show that face images artificially degraded through our generative approach are more realistic than those obtained with other statistical changes. Key disparities across gender and ethnic groups exist and urge timely interventions. Source code: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://cutt.ly/FChLfOC</uri>

Parallel deep learning architecture with customized and learnable filters for low-resolution face recognition

Parallel deep learning architecture with customized and learnable filters for low-resolution face recognition

Low Resolution Face Recognition Algorithm Based on MB-LBP

Low Resolution Face Recognition Algorithm Based on MB-LBP

Low resolution face recognition algorithm based on MB-LBP

Low resolution face recognition algorithm based on MB-LBP

Low-Resolution Face Recognition and Sports Training Action Analysis Based on Wireless Sensors

This paper constructs a low-resolution model for face recognition and sports training actions based on wireless sensors. The model obtains the distribution of the information size in the face image by calculating the image entropy value, and assigns different weights according to the size of the information to perform face recognition calculation, so that the original module-based algorithm is simply based on image segmentation into one based on entropy. The size of the value is divided into blocks, which solves the problem of computational quantification of category information. In the test stage, the traditional orthogonal matching pursuit algorithm is used to solve the coding coefficients, and the excellent classification and recognition results are obtained by calculating the intra-class matrix of the face image and the inter-class matrix of the sports training action image. Methods that perform well on classification problems further improve face recognition rates. The specific processing process is to add Gaussian noise, salt and pepper noise to the input face image and reduce the size of the face image in the input image, so that the improved algorithms are improved. The experimental results show that the high-efficiency resolution sensing technology is used to learn the sports training actions corresponding to the two modalities, and the matrix coefficient between the obtained high-resolution modal and low-resolution modal images reaches 0.971, and the iteration rate is improved by 71.5%, effectively promoting the high recognition rate of faces and actions.

Self-adjusting multilayer nonlinear coupled mapping for low-resolution face recognition

In practical surveillance scenes, effective recognition of low-resolution (LR) images is an urgent issue. At present, the improvement of methods for LR image recognition tasks mainly focuses on how to effectively extract information from LR images. This paper explores a novel self-adjusting multilayer nonlinear coupled mapping (SMNCM) network. The proposed SMNCM includes two main components, the multilayer nonlinear coupled mapping (MNCM) module, which is the feature extractor, and the self-adjusting feature fusion (SaFF) module. In the MNCM module, the LR template image set and multiple different high-resolution template image sets are used to construct a pyramid-like multicoupled mapping feature extraction network. In the SaFF module, the fusion weights of multiple features and the other parameters are collaboratively learned for the purpose of correct classification by the gradient learning network. To achieve this goal, a new gradient optimization model for fusion weights and other parameters is constructed. The features fused by the learned weights and the multiple features of the MNCM module are more suitable for classification tasks. Experimental results indicate that on the five databases, our proposed method outperforms the state-of-the-art approaches for LR face recognition. The ablation experiment suggests that each module in the proposed method can contribute to improving the LR image recognition performance.

A transformer-based low-resolution face recognition method via on-and-offline knowledge distillation

It has been widely noticed the performance of algorithms for high-resolution face recognition (HRFR) degrades significantly for low-resolution face recognition (LRFR). In this paper, we discover the main source of this performance degradation comes from the human-defined inductive bias of CNN, which constrains the model to absorb effective information and leads to overfitting. To overcome the shortcoming, for the first time, we adopt a transformer-based DNN algorithm named DeiT to accomplish LRFR tasks. On the other hand, we further borrow the form of knowledge distillation. Traditional knowledge distillation network for LRFR sets the student model to be simpler than or the same as the teacher model while using the teacher model off-the-shelf, leading to the model capacity gap. Instead, we fuse an online network into the original parameter-fixed teacher model to learn how to transfer knowledge. The final “knowledge” is the sum fusion of the outputs of both the teacher model and the student model.Experiments show that even without training on LR face images, the performance of our model can be comparable to recent baselines delicately designed for LRFR tasks on certain LR face datasets. After the entire training process, in either real-world LR datasets, artificially down-sampled datasets, or generated LR face datasets, our method performs favorably against the state-of-the-art (SOTA) algorithms.

Adaptive random down-sampling data augmentation and area attention pooling for low resolution face recognition

State-of-the-art face recognition method have achieved satisfactory performance on various face datasets, but limited to high-resolution (HR) faces. In real scenes, with affects from camera quality or distance etc., low quality face images are often obtained in surveillance scenes, which leads to poor performance of general face recognition approaches. In this work, we explored the reasons why low-resolution (LR) images cannot be coped with the conventional models, and proposed the optimization methods. The main contribution of this work includes two aspects: (1) We proposed an adaptive random down-sampling data augmentation method with the corresponding loss function and training strategy. For the network training, we mapped the face images with different resolutions to the same embedding space, and then explicitly separated different classes of LR and gathered the same class of HR and LR. (2) A pooling module based on attention mechanism, area attention pooling, was proposed. After pooling, the fine-grained information of the identity is retained, and the redundant information of the non-identity is removed. We conducted a series of visualization experiments to prove the effectiveness of our method. We evaluated our methods on the real-world LR face dataset Surveillance Cameras Face Database (SCface) and the public face evaluation dataset Labeled Faces in the Wild (LFW). Compared with state-of-the-art methods, the proposed method achieves satisfactory performance at LR images on SCface and LFW dataset.

Low-resolution face recognition based on feature-mapping face hallucination

The image recognition approaches based on Convolutional Neural Network (CNN) have already achieved tremendous performance on super-resolution face images. However, there exist several challenges in face recognition field. For example, in very low-resolution face recognition (LRFR) environments, the accuracy will drop drastically. To address the issue, this paper proposes a novel face hallucination and recognition model for low resolution face images ground on feature-mapping. In the proposed model, a new loss function named identity-aware loss is also proposed. The proposed loss function is combined with the feature loss and image-content loss to jointly train models. The proposed model is evaluated on Labeled Faces in the Wild (LFW) dataset, which compared to progressive competing models. A large number of experimental results indicate that this model observably enhances the performance of recognition particularly when face images are very low-resolution. In addition, our model can perform high resolution face image reconstruction which is comparable to advanced approaches based on super-resolution in the field of visual quality, and achieve identity preservation of corresponding low resolution probe image.

Multi-Scale Identity-Preserving Image-to-Image Translation Network for Low-Resolution Face Recognition

State-of-the-art deep neural network models have reached near perfect face recognition accuracy rates on controlled high-resolution face images (especially Caucasian faces). However, their performance is drastically degraded when they are tested with very low-resolution face images. This is particularly critical in surveillance systems, where a low-resolution probe image is to be matched with high-resolution gallery images. Super-resolution techniques aim at producing high-resolution face images from low-resolution counterparts. While they are capable of reconstructing images that arevisually appealing, the identity-related information is not preserved. Here, we propose an identity-preserving end-to-end image-to-image translation deep neural network which is capable of super-resolving very low-resolution faces to their high-resolution counterparts while preserving identity-related information. We achieved this by training a very deep convolutional encoder-decoder network with a symmetric contracting path between corresponding layers. This network was trained with a combination of a reconstruction andan identity-preserving loss, on multi-scale low-resolution conditions. Extensive quantitative evaluations of our proposed model demonstrated that it outperforms competing super-resolution and low-resolution face recognition methods on natural and artificial low-resolution face data sets and even unseen identities.

Adapted discriminative coupled mappings for low-resolution face recognition with one-shot

Low-resolution face recognition with one-shot is a prevalent problem encountered in law enforcement, where it generally requires to recognize the low-resolution face images captured by surveillance cameras with the only one high-resolution profile face image in the database. The problem is very tough because the available samples is quite few and the quality of unknown images is quite low. To effectively address this issue, this paper proposes Adapted Discriminative Coupled Mappings (AdaDCM) approach, which integrates domain adaptation and discriminative learning. To achieve good domain adaptation performance for small size dataset, a new domain adaptation technique called Bidirectional Locality Matching-based Domain Adaptation (BLM-DA) is first developed. Then the proposed AdaDCM is formulated by unifying BLM-DA and discriminative coupled mappings into a single framework. AdaDCM is extensively evaluated on FERET, LFW, and SCface databases, which includes LR face images obtained in constrained, unconstrained, and real-world environment. The promising results on these datasets demonstrate the effectiveness of AdaDCM in LR face recognition with one-shot.

Deep Rival Penalized Competitive Learning for low-resolution face recognition

Current face recognition tasks are usually carried out on high-quality face images, but in reality, most face images are captured under unconstrained or poor conditions, e.g., by video surveillance. Existing methods are featured by learning data uncertainty to avoid overfitting the noise, or by adding margins to the angle or cosine space of the normalized softmax loss to penalize the target logit, which enforces intra-class compactness and inter-class discrepancy. In this paper, we propose a deep Rival Penalized Competitive Learning (RPCL) for deep face recognition in low-resolution (LR) images. Inspired by the idea of the RPCL, our method further enforces regulation on the rival logit, which is defined as the largest non-target logit for an input image. Different from existing methods that only consider penalization on the target logit, our method not only strengthens the learning towards the target label, but also enforces a reverse direction, i.e., becoming de-learning, away from the rival label. Comprehensive experiments demonstrate that our method improves the existing state-of-the-art methods to be very robust for LR face recognition.

An Implicit Identity-Extended Data Augmentation for Low-Resolution Face Representation Learning

Low-resolution (LR) face recognition (LRFR) tackles <i>tiny face images</i> detected from real-world surveillance camera footage, which are unconstrained and generally poor in quality. Owing to the absence of a million-scale labeled LR face dataset, <i>identity-invariant</i> data augmentation (DA) transformations such as flipping, rotation, rescaling, etc., are applied to inflate the effective training examples with respect to the source identities for representation learning. Unfortunately, the identity-invariant property incurs additional intra-class disparity that impairs generalization performance. In this paper, we put forward a new means of DA strategy, termed <i>identity-extended</i> DA, that satisfies both <i>affinity</i> and <i>diversity</i> requirements essential to DA. We instantiate an implicit identity-extended augmentation network, or simply IDEA-Net, to realize the proposed identity-extended DA for LRFR. More specifically, training an IDEA-Net instance augments the small-scale LR (query) face dataset with identity-extended (auxiliary) face examples implicitly in the representation space. We also introduce a calibrator to regulate the disordered representation space by refining the intra-class compactness and the inter-class separation. This diminishes the distribution shift between the original and the augmented examples (affinity) and increases the learning complexity (diversity). We substantiate that IDEA-Net renders a high affinity and diversity representation space. On the other hand, our experimental results on three real-world LR face datasets demonstrate that IDEA-Nets outperform the baselines and other counterparts trained without leveraging the identity-extended examples for LRFR.

Multiscale parallel deep CNN (mpdCNN) architecture for the real low-resolution face recognition for surveillance

Images from the surveillance networks are being extensively used for the purpose of monitoring and criminal investigations. However, it is often difficult to recognize faces using the data from the surveillance networks because the resolution of the face in the images captured is too low. Further, the low-resolution images have varying magnitude of facial feature content because of wide variations in illumination, pose, resolution and the distance from which the image is captured. Also, a single face recognition solution is not able to recognize faces efficiently in both high and low-resolution images. Wide variations in facial feature content in high and low-resolution images causes difficulty in classification of the features by a single model for the purpose of face recognition. We present a Deep-CNN based architecture called mpdCNN to solve the problem of face recognition in low as well as high-resolution images with high accuracy and robustness. Our proposed architecture mpdCNN gives 88.6% accuracy on the SCface database which is an impressive improvement over the state-of-the-art algorithms. We also achieved an accuracy of above 99% on normal to high-resolution databases for face recognition.

Low-resolution face recognition in resource-constrained environments

• A LR face recognition model for resource-constrained environments is proposed. • Our model is lightweight and capable of being effectively trained on small data. • Our model adopts PixelHop++ which is designed based on the SSL principle. • Active learning is incorporated to minimize the required labeled training data. • We show the competitive performance of our model compared with state-of-the-art. Although Deep Neural Networks (DNNs) have achieved tremendous success in the face recognition task, utilizing them in resource-constrained environments with limited networking and computing is challenging. Such environments often demand a small model capable of being effectively trained on a small number of labeled training data, with low training complexity, and low-resolution input images. To address these challenges, we adopt an emerging machine learning methodology called Successive Subspace Learning (SSL) to propose LRFRHop, a high-performance data-efficient low-resolution face recognition model for resource-constrained environments. SSL offers an explainable non-parametric feature extraction submodel that flexibly trades the model size for the verification performance. Its training complexity is significantly lower than DNN-based models since it is trained in a one-pass feedforward manner without backpropagation. Furthermore, active learning can be conveniently incorporated to reduce the labeling cost. We demonstrate the effectiveness of LRFRHop by conducting experiments on the LFW and the CMU Multi-PIE datasets.

Special issue on deep learning for emerging big multimedia super-resolution

The main goal of the super-resolution (SR) is to restore a visually pleasing high-resolution image using a low-resolution image of video sequence. The higher resolution image is composed of higher pixel density with fine and precise details as compared with the low-resolution images or video. The majority of the applications, such as video surveillance, ultra-high-definition TV, low-resolution face recognition and remote sensing imaging are based on super-resolutions. Thus benefiting from the broader spectrun of these applications, the super-resolution has attracted massive interest form both acemedia and industry. And currently, a most active research field in today’s era.